Tool

A Tool declares a capability the agent can invoke on its own when it deems necessary. It connects to the tools port of an AI Agent and you can connect several.

The powerful part: a tool can run any of the 600+ nodes in Floogos (generate images, call an API, query an ERP, read an Excel file…), a whole workflow, or a registered tool.

Configuration

Field	Type	Required	Description
tool_name	text	Yes	The function name the LLM will see (e.g. get_weather). No spaces.
tool_description	textarea	Yes	What it does. The LLM decides when to use it by reading this.
tool_schema	json	No	JSON schema of the parameters (JSON Schema).
tool_kind	select	Yes	node, workflow or registered.
node_type	text	Cond.	If tool_kind = node: the node type (e.g. openaiImages, http).
node_config	json	No	Base config of the node (merged with the LLM’s args).
workflow_id	text	Cond.	If tool_kind = workflow: the workflow id.
tool_id	text	Cond.	If tool_kind = registered: the id in agent_tools.
attach_to	text	Yes*	Agent alias.

A good description is key

The model chooses which tool to use by reading its tool_description. Be explicit: what it does and when to use it. A vague description = a tool that isn’t used (or is used wrong).

Example 1 — A node as a tool (generate images)

1. Tool:
   • tool_name: generate_image
   • tool_description: “Generates an image from a description. Use it when the user asks to create, draw or illustrate something.”
   • tool_kind: node
   • node_type: openaiImages
   • tool_schema:

     {
       "type": "object",
       "properties": {
         "prompt": { "type": "string", "description": "Image description" }
       },
       "required": ["prompt"]
     }

2. Connect its output to the tools port.

For “draw me an astronaut cat”, the LLM calls generate_image with { prompt: "astronaut cat" }, the agent runs openaiImages and returns the result.

Example 2 — A workflow as a tool

When the capability is a multi-step process, wrap it in a workflow and expose it as a single tool.

1. Tool — tool_name: create_customer, tool_kind: workflow, workflow_id: 142 (the id of your onboarding workflow).
2. tool_description: “Creates a new customer. Requires name and email.”
3. tool_schema: define name and email as parameters.

The agent runs workflow 142 with the data the LLM extracted from the request.

Example 3 — Query an ERP

tool_kind: node, node_type: finnegansGet, and in node_config the fixed credential/entity; the LLM provides the filter. This lets the agent read data from your ERP during the conversation.

Caution

If the tool fails with “Node ‘X’ not found”, check that node_type is the node’s real type (e.g. the CSV node is csvparce). If the node needs credentials, pass them in node_config.

See also

• AI Agent (orchestrator) · Skill (for tool gating)
• Examples and recipes · Common problems
• Agent library — ready-made skills, tools, prompts and kits to insert into your agent.